NL80211_CMD_ROAM indication is scheduled via a kernel work queue, while
QCA_NL80211_VENDOR_SUBCMD_KEY_MGMT_ROAM_AUTH is a vendor event from the
driver. Thus, a race condition can exist wherein the vendor event is
received prior to the NL80211_CMD_ROAM indication.
The processing of this vendor event depends on the NL80211_CMD_ROAM
indication to update the roamed BSS/BSSID information and thus the out
of sequence processing of these events would result in not updating the
right BSS information.
This commit adds a workaround to hold the pending
QCA_NL80211_VENDOR_SUBCMD_KEY_MGMT_ROAM_AUTH event for up to 100 ms in
case NL80211_CMD_ROAM is not received first.
Signed-off-by: Purushottam Kushwaha <pkushwah@codeaurora.org>
Support possible band combinations of 2.4 GHz, 5 GHz, and 6 GHz with
QCA_WLAN_VENDOR_ATTR_SETBAND_MASK attribute. Ensure backwards
compatibility with old drivers that are using
QCA_WLAN_VENDOR_ATTR_SETBAND_VALUE attribute and supporting only 2.4 GHz
and 5 GHz bands.
Signed-off-by: Veerendranath Jakkam <vjakkam@codeaurora.org>
The kernel requires indication of DFS handler residing in user space
(NL80211_ATTR_HANDLE_DFS) to enable DFS channels.
Signed-off-by: Peter Oh <peter.oh@bowerswilkins.com>
mac80211 does not allow mgmt tx to use off channel on
DFS channels in non-ETSI domain, because it will invalidate
CAC result on current operating channel.
(mac80211 commit: 34373d12f3cbb74960a73431138ef619d857996f)
Hence don't set offchanok for mgmt tx in case of DFS channels
in non-ETSI.
Signed-off-by: Peter Oh <peter.oh@bowerswilkins.com>
Use NL80211_ATTR_SAE_PWE attribute to indicate the sae_pwe value
to the driver during the NL80211_CMD_START_AP and NL80211_CMD_CONNECT
in WPA3-Personal networks which are using SAE authentication.
Signed-off-by: Rohan Dutta <drohan@codeaurora.org>
The newer kernel versions enforce strict netlink attribute policy
validation and will cause cfg80211 to reject vendor commands with
NL80211_ATTR_VENDOR_DATA if NLA_F_NESTED attribute is not set but
if the vendor command is expecting nested data within
NL80211_ATTR_VENDOR_DATA attribute.
Most of the earlier instances were addressed by adding NLA_F_NESTED
flag in nla_nest_start(). This commit addresses the remaining
instance in which NL80211_ATTR_VENDOR_DATA is populated using data
set by user through the control interface.
Enhance the control interface VENDOR command to indicate whether the
vendor subcommand uses nested attributes within NL80211_ATTR_VENDOR_DATA
attribute or not.
Set NLA_F_NESTED flag for existing QCA vendor commands which use nested
attributes within the NL80211_ATTR_VENDOR_DATA attributes so that the
old frameworks implementations for already existing commands work
without any issues.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
When nlmsg allocation fails, nl80211_drv_msg() returns NULL and the call
to send_and_recv_msgs_owner() from nl80211_leave_ibss() could have ended
up dereferencing a NULL pointer. Fix this by make
send_and_recv_msgs_owner() more consistent with other send_and_recv*()
cases that check msg == NULL internally.
Fixes: 12ea7dee31 ("nl80211: Use nl80211 control port for receiving EAPOL frames")
Signed-off-by: Pooventhiran G <pooventh@codeaurora.org>
Add support to receive and process MSCS Response frames from the AP and
indicate the status to upper layers.
Signed-off-by: Vinita S. Maloo <vmaloo@codeaurora.org>
Use the "tmp_disallow" name more consistently so that both the core
wpa_supplicant functionality (struct wpa_bss_tmp_disallowed) and the
wpa_driver_ops callback have more similar names.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
The NL80211_ATTR_VLAN_ID attribute expects non-zero values, but vlan_id
with value 0 has been set in VLAN offload case. Due to this, station
connection failure is observed if the driver advertises VLAN_OFFLOAD
support:
nl80211: NL80211_ATTR_STA_VLAN (addr=8c:fd:f0:22:19:15 ifname=wlan0
vlan_id=0) failed: -34 (Result not representable)
wlan0: STA 8c:fd:f0:22:19:15 IEEE 802.11: could not bind the STA
entry to vlan_id=0
Fix this by setting only non-zero values.
Fixes: 0f903f37dc ("nl80211: VLAN offload support")
Signed-off-by: Seevalamuthu Mariappan <seevalam@codeaurora.org>
Check if nl80211 control port TX status is available in the kernel and
enable control port TX if so. With this feature, nl80211 control path is
able to provide the same feature set as nl80211 (management) + AF_PACKET
socket (control) before.
For debugging and testing, this can explicitly be disabled with
the driver parameter control_port_ap=0.
Signed-off-by: Markus Theil <markus.theil@tu-ilmenau.de>
The kernel commit "mac80211: support control port TX status reporting"
seems to be delivering the TX status events for EAPOL frames over
control port using NL80211_CMD_FRAME_TX_STATUS due to incorrect check on
whether the frame is a Management or Data frame. Use the pending cookie
value from EAPOL TX operation to detect this incorrect behavior and
redirect the event internally to allow it to be used to get full TX
control port functionality available for AP mode.
Signed-off-by: Jouni Malinen <j@w1.fi>
Allow custom ack handler to be registered and use the ext ack handler
for TX control port to fetch the cookie information. If these cookies
are not supported by the current kernel, a value of 0 is returned.
Signed-off-by: Markus Theil <markus.theil@tu-ilmenau.de>
This is a preliminary patch for using extack cookies for TX control port
handling. Custom ack handler arguments for send_and_recv() and friends
is introduced therefore. This commit does not actually use the provided
values, i.e., that will be added in a separate commit.
Signed-off-by: Markus Theil <markus.theil@tu-ilmenau.de>
wpa_supplicant registered to process the Radio Measurement Request
frames unconditionally. This would prevent other location based
applications from handling these frames. Enable such a use case by
allowing wpa_supplicant to be configured not to register to process
these frames. This can now be done by adding "no_rrm=1" to the
driver_param configuration parameter.
In addition, wpa_driver_nl80211_init() does not have the provision to
take driver_params. Hence, resubscribe again with cfg80211 when this
driver parameter "no_rrm=1" is set after the initial setup steps.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Use a shared wpa_alg_bip() function for this and fix the case in
nl_add_key() to cover all BIP algorithms. That fix does not change any
behavior since the function is not currently used with any BIP
algorithm, but it is better to avoid surprises should it ever be needed
with IGTK.
Signed-off-by: Jouni Malinen <j@w1.fi>
IGTK is the key that is used a BIP cipher. WPA_ALG_IGTK was the
historical name used for this enum value when only the AES-128-CMAC
based BIP algorithm was supported. Rename this to match the style used
with the other BIP options.
Signed-off-by: Jouni Malinen <j@w1.fi>
Commit dd74ddd0df ("nl80211: Handle AKM suite selectors for AP
configuration") added warning log message "nl80211: Not enough room for
all AKM suites (num_suites=X > NL80211_MAX_NR_AKM_SUITES)" which in some
cases fills logs every 3 seconds, so fix this by increasing the log
message level to debug.
Reported-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
Ref: https://patchwork.ozlabs.org/project/openwrt/patch/20200504130757.12736-1-ynezz@true.cz/#2429246
Fixes: dd74ddd0df ("nl80211: Handle AKM suite selectors for AP configuration")
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
Linux bridging code does not allow a station mode WLAN interface in a
bridge and this prevents the AP mode scan workaround from working if the
AP interface is in a bridge and scanning can be only done by moving to
STA mode. Extend this workaround to remove the interface from the bridge
temporarily for the duration of the scan, i.e., for the same duration as
the interface needs to be moved into the station.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
When hostapd is started without beaconing (start_disabled=1), Public
Action frame transmission command through nl80211 needs to allow
offchannel operations regardless of the operating channel configuration.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
The driver can advertise supported AKMs per wiphy and/or per interface.
Populate per interface supported AKMs based on the driver advertisement
in the following order of preference:
1. AKM suites advertised by NL80211_ATTR_IFTYPE_AKM_SUITES
2. AKM suites advertised by NL80211_ATTR_AKM_SUITES
If neither of these is available:
3. AKMs support is assumed as per legacy behavior.
In addition, extend other driver interface wrappers to set the
per-interface values based on the global capability indication.
Signed-off-by: Veerendranath Jakkam <vjakkam@codeaurora.org>
Update frame processing registration for DPP Public Action frames to
include multicast reception when in DPP listen mode and the driver
indicates support for this type of explicit request.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
This has no changes to the function itself, i.e., it is just moved to a
location that is above the first caller.
Signed-off-by: Jouni Malinen <j@w1.fi>
drv->eapol_sock is used only for receiving EAPOL frames in AP mode, so
it is not needed when using control port for EAPOL frame RX.
Signed-off-by: Jouni Malinen <j@w1.fi>
Since this nl80211 mechanism for sending EAPOL frames does not currently
support TX status notification, disable it by default of AP mode where
the Authenticator state machine uses those notifications to optimize
retransmission. The control port TX can be enabled for AP mode with
driver param control_port_ap=1.
Signed-off-by: Jouni Malinen <j@w1.fi>
Use and/or set socker owner where necessary to allow EAPOL frames to be
received using the nl80211 control port. This is done when the driver
indicates support for the control port without previous hardcoded
reception of RSN preauth frames.
Use methods which set or use the connection owner nl_sock * where
necessary. Initial operations need to register with the SOCKET_OWNER
attribute set (e.g., connect for STA mode). Final operations need to use
the socket which holds the owner attribute (e.g., disconnect for STA
mode).
Signed-off-by: Markus Theil <markus.theil@tu-ilmenau.de>
This commit creates a connect nl_sock * for every bss unconditionally.
It is used in the next commit for nl80211 control port RX.
Signed-off-by: Markus Theil <markus.theil@tu-ilmenau.de>
This is needed since the initial implementation of the control port
mechanism in the kernel mixed in RSN pre-authentication ethertype
unconditionally (and IMHO, incorrectly) into the control port.
Signed-off-by: Jouni Malinen <j@w1.fi>
The only l2_packet use within this file was removed by commit
d12dab4c6f ("nl80211: Use non-receiving socket for EAPOL TX").
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
The new driver param full_ap_client_state=0 can be used to test
functionality with the driver capability for full AP client state being
forced to be disabled.
Signed-off-by: Jouni Malinen <j@w1.fi>
Drivers that trigger roaming need to know the lifetime and reauth
threshold time of configured PMKSA so that they can trigger full
authentication to avoid unnecessary disconnection. To support this, send
dot11RSNAConfigPMKLifetime and dot11RSNAConfigPMKReauthThreshold values
configured in wpa_supplicant to the driver while configuring a PMKSA.
Signed-off-by: Veerendranath Jakkam <vjakkam@codeaurora.org>
hostapd will trigger EDMG auto channel selection by setting
QCA_WLAN_VENDOR_ATTR_ACS_EDMG_ENABLED. The 60 GHz driver will be
called to start an auto channel selection and will return the
primary channel and the EDMG channel.
Signed-off-by: Noam Shaked <nshaked@codeaurora.org>
If the kernel rtm_newlink or rtm_dellink send the maximum length of
ifname (IFNAMSIZ), the event handlers in
wpa_driver_nl80211_event_rtm_addlink() and
wpa_driver_nl80211_event_rtm_dellink() did not copy the IFLA_IFNAME
value. Because the RTA_PAYLOAD (IFLA_IFNAME) length already includes the
NULL termination, that equals the IFNAMSIZ.
Fix the condition when IFNAME reach maximum size.
Signed-off-by: Ouden <Ouden.Biz@gmail.com>
Add key configuration parameters needed to support Extended Key ID with
pairwise keys. Add a driver capability flag to indicate support forusing
this.
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
The return value from nl80211_send_monitor() is not suitable for use
with strerror(). Furthermore, nl80211_send_monitor() itself is printing
out a more detailed error reason.
Signed-off-by: Jouni Malinen <j@w1.fi>
send_and_recv_msgs() returns a negative number as the error code and
that needs to be negated for strerror().
Fixes: 8759e9116a ("nl80211: Control port over nl80211 helpers")
Signed-off-by: Jouni Malinen <j@w1.fi>
wpa_tdls_set_key() did set the key_id to -1 to avoid a useless
NL80211_CMD_SET_KEY call that the updated nl80211 driver no longer
carries out. Remove the no longer required workaround.
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
Migrate nl80211 driver to key_flag API and add additional sanity checks.
I'm still not sure why we install unicast WEP keys also as default
unicast keys. Based on how I understand how mac80211 handles that it
should be pointless. I just stuck to how we do things prior to the patch
for WEP keys to not break anything. After all other drivers may need it.
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
Always report an error when NL80211_CMD_SET_KEY can't set a key to
default.
The old ioctl-based design used a single command to add, set, and delete
a key and had to ignore ENOENT for key deletions. It looks like that
special handling was also ported for NL80211_CMD_NEW_KEY and
NL80211_CMD_SET_KEY instead only for NL80211_CMD_DEL_KEY.
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
Calling NL80211_CMD_NEW_KEY with NL80211_KEY_DEFAULT_TYPES attributes is
pointless. The information is not expected and therefore the kernel
never forwards it to the drivers. That attribute is used with
NL80211_CMD_SET_KEY.
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
Allow to abort key installations with different error codes and fix one
misleading return code.
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
Turns out we are sometime providing a seq when deleting the key. Since
that makes no sense on key deletion let's stop forwarding that to the
driver at least.
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
PASN authentication can be performed while a station interface is
connected to an AP. To allow sending PASN frames while connected, extend
the send_mlme() driver callback to also allow a wait option. Update the
relevant drivers and wpa_supplicant accordingly.
hostapd calls for send_mlme() are left unchanged, since the wait option
is not required there.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
If the driver requires external SAE authentication, it would result in
registration for all Authentication frames, so even non-SAE
Authentication frames might be forwarded to user space instead of being
handled internally. Fix this by using a more strict match pattern,
limiting the registration to the SAE authentication algorithm only.
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
